Credit scoring and the credit-underserved population

Credit scores play an increasingly important role in consumers' lives. Not only are they used to determine whether or not an individual has access to consumer credit and at what price; they also influence the price of insurance, the ability to rent an apartment, and the hiring decisions of employers, among other things.

Credit scoring helps determine the financing options available to the estimated 160 million Americans who participate in the mainstream credit market.^1/ Consumers with high credit scores reap the benefits of paying low interest rates and having an array of credit options. For those with low credit scores, the options are fewer and costlier. However, if individuals with bad credit histories take steps to improve their scores, opportunities to access affordable credit will become available.

But what about the other American consumers? The millions of people who have little or no history at the major credit bureaus? These consumers, whom the financial services industry describes as underscored or credit-underserved, are meeting many of the same financial obligations that the majority of consumers meet. For example, credit-underserved consumers make payments for rent, utilities, or other necessities. But since the mainstream credit market does not collect complete data about those sorts of payments, mainstream lenders often have too little information about the credit-underserved to efficiently extend loans to them.

When these consumers need to borrow money, the options available to them are limited and expensive. Credit-underserved consumers often end up paying high rates, fees, and down payments. For example, many in the credit-underserved market turn to payday lenders and check cashing services that charge effective interest rates as high as 500 percent.^2/

Fortunately, there are new means of credit scoring in development that can help underserved consumers enter the mainstream American credit market. A movement is under way to collect and score alternative data that reflect the many payments credit-underserved individuals routinely make for insurance, utilities, and other products and services. A survey of the movement reveals that the use of alternative data and scoring offers promise, but barriers remain.

A huge, diverse group

Who are the credit-underserved? Estimates of the group's size range from 35 million to 70 million adults, depending on the source of information and how the market is defined (see the table below). Research shows that the credit-underserved market includes many immigrants who may have little or no credit history from their home countries; young adults who have had little time to build a credit history; recently divorced or widowed individuals who, having previously relied on their spouses to manage the household finances, have never borrowed money in their own names; and groups that are culturally averse to credit use, including retirees and ethnic groups that distrust banks and other credit granters.^3/

The credit-underserved fall into two main categories: no-file consumers, or consumers who have no credit history on file with the credit bureaus; and thin-file consumers, or consumers about whom credit bureaus have very little information. Thin credit files might contain only derogatory data that do not provide a balanced representation of a consumer's creditworthiness. For example, a thin file could include a record of missed payments for telephone service, but omit any record of regular, on-time payments for other services.

The credit-underserved market: Estimates and definitions Data Source Estimated Size of Credit-Underserved Market Data Source's Definition of "Credit-Underserved" Experian 35 million adults Not "credit active" Fair Isaac 54 million adults Having no credit files that can be scored, either because of no credit history (22 million adults) or thin credit files (32 million adults) National Credit Reporting Association 70 million adults Having either no credit score or a lower credit score than their financial history and payment potential warrant Source: Katy Jacob and Rachel Schneider, Market Interest in Alternative Data Sources and Credit Scoring, The Center for Financial Services Innovation, December 2006.

A system evolves

How did we arrive at a point where national credit reporting and scoring systems have such power over consumers' lives?

Prior to the nineteenth century, lenders in the U.S. used informal, locally gathered information to subjectively evaluate the creditworthiness of borrowers. Partly as a result, consumer lending was limited, expensive, and not always competitive. In the nineteenth century, formal credit reporting systems took root when groups of retail merchants came together to share information about their customers' financial habits and payment histories. These efforts grew into merchant associations, which later morphed into small credit bureaus, also called credit reporting agencies (CRAs).

By the late twentieth century, the advent of computerization enabled CRAs to efficiently amass enormous amounts of data, and lending institutions began to rely on CRAs as a major source of underwriting information. Underwriting, the process used to analyze and predict how a borrower manages credit obligations, was once performed manually and took between 30 and 60 days. Technological advances ushered in the era of automated underwriting, in which computerized systems analyze information from loan applications and arrive at near-instantaneous, logic-based decisions to approve or deny loans.

At the same time that CRAs and automated underwriting were evolving, two mathematicians, William Fair and Earl Isaac, began work that would help lenders better leverage the data held by CRAs. Fair and Isaac developed models that predicted lending risk by determining which factors were good and reliable predictors of a consumer's future debt-payment performance. Their models assigned numerical scores to indicate the creditworthiness of each borrower.^4/ The Fair Isaac Corporation and its FICO (Fair Isaac Corporation) score would evolve to become the most well known and widely used credit score in the U.S.

While the emerging credit reporting and scoring system made underwriting more efficient, it posed some risks to borrowers. Early CRAs lacked safeguards for ensuring the privacy of the information they held. They collected data about the negative aspects of consumers' credit histories, such as delinquencies, defaults, and bankruptcies, while minimizing information about on-time payments. Also, they recorded "lifestyle" information, gathered from newspapers and other sources, that included personal details such as sexual orientation, drinking habits, and cleanliness. Moreover, consumers were blocked from viewing or correcting their files.

Controversy over the effects of this system on borrowers led to a congressional inquiry and passage of the Fair Credit Reporting Act (FCRA) in 1971. FCRA established a framework to protect privacy and promote accuracy in credit reporting. The act gave consumers the right to view, dispute, and correct their records, and the CRAs began to systematically collect and report information on consumers' positive financial histories. In 2003, FCRA was amended to allow consumers to request and obtain a free credit report and, for a "fair and reasonable" fee, a notice of their credit score.^{5/ 6/} (For more on obtaining a free credit report, visit www.annualcreditreport.com.)

Credit reporting today

Today, CRAs and credit scoring models form the foundation of our credit system. The financial habits of most American consumers are monitored by one or more of the three large, national CRAs: Equifax, Experian, and TransUnion. Every month, creditors transmit more than 4.5 billion pieces of data to one or more of these organizations. The CRAs compile the data into the credit records of individual consumers. Lenders then use these records to perform the automated underwriting that has made the extension of instant credit possible.

Information reported to CRAs includes consumers' account numbers, the various types of credit held (mortgage loans, credit card loans, car loans, etc.), outstanding balances, and collection actions. Typically, creditors provide this information to CRAs in exchange for access to the credit records of other borrowers. In this way, creditors can efficiently target and market their credit products to more customers. In addition to information creditors provide, public records like court judgments, claims of overdue child support, bankruptcies, foreclosures, and liens are also compiled by CRAs and included in credit reports. (For a detailed description of the information credit reports contain, see the February 2003 Federal Reserve Bulletin article listed in the "For more information" box below.)

Upon lender request, CRAs calculate credit scores, such as the FICO score, from the information they collect in an individual borrower's credit report. The scoring systems analyze this information and award points for each factor that predicts a high probability of an individual borrower repaying his or her debts on time. The total number of points—the credit score—is a tool for predicting how creditworthy a person is.^7/

The good, the bad, and the underserved

As the efficiency of the reporting and scoring system has increased, the applications for credit reports and scores have expanded dramatically. According to a Brookings Institution report, credit reports and scores are not only used to decide if a consumer can borrow money for a home or car. Businesses also use credit scores to evaluate prospective apartment renters; assess job applicants; and price mortgages, deposits for utility services, and various types of insurance.^8/

As noted earlier, the expanding influence of credit scoring on consumers' lives can have positive or negative results for an individual, depending on his or her financial history and habits. The robust reporting and scoring system in the U.S. has facilitated a credit market that allows lenders to screen potential borrowers efficiently, extend credit quickly, and manage risk. This development has historically provided a benefit to U.S. consumers—particularly consumers who traditionally would have had little or no access to credit—beginning with a vast expansion of the credit card market in the late 1980s. Later on, credit scoring made mortgage underwriting faster and cheaper, which helped broaden access to mortgages and homeownership.

The growing availability of credit has also expanded the resources available to new entrepreneurs launching businesses, and has given many families access to the funds they need to "smooth over" periods of financial challenge.^9/ At the same time, competition among lenders for individuals with solid credit histories has reduced the price of credit for those consumers.^10/

Of course, there is also a downside to making credit more readily available. For example, a relatively high percentage of first-time borrowers will default on their credit cards, mortgages, and other loans. However, provided appropriate underwriting standards are maintained, the benefits of more efficient and objective underwriting and broader access to credit should outweigh the downside.

Whatever the effects of the credit reporting and scoring system, it can be argued that having a credit history—blemished or unblemished—is better than not having one. A consumer who repairs a blemished credit history can, through the power of automated underwriting, gain access to affordable sources of credit. In contrast, a credit-underserved consumer might never gain access to affordable credit, because his or her credit history is too scant to be processed by an automated underwriting system. These individuals are left with limited access to credit and the potential asset-building opportunities it offers. And, their access to necessities like rental housing, jobs, and home energy utilities is increasingly constrained.

While credit-underserved consumers have participated little or not at all in the traditional credit market, they are meeting other types of payment obligations. Like other Americans, credit-underserved individuals make monthly payments for rent and utility services. Many also regularly pay for insurance, savings plans, childcare, health care, or interest and principal on alternative loans, such as payday loans.

This final point has led many in the community development and financial services industries to look at the possibility of establishing or augmenting credit histories for credit-underserved consumers by collecting and scoring data related to alternative, noncredit-based payment obligations.

Alternatives in the works

The use of alternative data and scoring to bring credit-underserved individuals into the mainstream credit market offers benefits for both lenders and borrowers. Many underserved borrowers would benefit from the opportunity to access credit more readily and at cheaper prices. Lenders would benefit from having the information they need to extend credit to a large and untapped market. Fair Isaac estimates that reaching just 3 percent of this market would put in play an additional $2.3 billion for mortgage lenders, $750 million for automobile lenders, and $113 million for credit card issuers.^11/

A variety of organizations, both inside and outside the traditional credit reporting and scoring system, are working to collect and analyze data from sources not currently reported to CRAs. Alternative data sources under consideration include payments for energy and telecommunications, auto liability and homeowner's insurance, rental housing, childcare, payday loans, health care, and certain types of retail payments (e.g., furniture rental data). The goal is to identify sources that can be used with credit scoring models to reliably predict the creditworthiness of credit-underserved individuals.

The Center for Financial Services Innovation (CFSI) has compiled and published information on alternative data collection and analysis efforts that are currently under way.^12/ Some highlights:

• Community Financial Services Association of America, the largest association of payday loan companies, is offering its customers the opportunity to have their repayment data reported to credit bureaus as part of a pilot project in select geographies.
• Fair Isaac has created the FICO Expansion Score, which uses historical data on utility and insurance payments, in combination with detailed information from loan applications, to calculate a credit score and recommend approval or rejection of a loan application.
• First American Corporation is marketing the Anthem (Assisting Nontraditional Homebuyers in Emerging Markets) report, which takes into account payment histories for rent, insurance, utility bills, and childcare expenses as well as traditional credit data to generate a credit report and score.
• Lexis-Nexis's RiskView uses more than 300 public record attributes, such as employment, previous addresses, and property and asset ownership, to verify consumers' identities and predict risk behavior.
• Link2Credit is a "credit decisioning platform" for wireless telecommunications carriers that uses historical phone payment records and other alternative data sources. Link2Credit is beginning to work with credit card companies and the mortgage industry.
• eFunds DebitBureau uses consumers' checking account histories as a basis for data about credit decisions.
• Payment Reporting Builds Credit (PRBC) is a new credit bureau that began by tracking self-reported rental payments. PRBC has expanded to include more general bill payment tracking and third-party verification.

The big question that these efforts are trying to answer is whether or not the alternative data and models used have strong predictive value. In other words, do the data accurately predict whether or not a borrower will repay his or her debts in a timely manner? Initial analysis suggests that the answer is yes.

A recent CFSI review of early test data on three of the projects listed above—FICO Expansion Score, RiskView, and Link2Credit—concluded that alternative credit scores can be generated for most individuals who lack traditional credit scores and that alternative scores have meaningful, predictive value for lenders who extend credit to individuals with little or no traditional credit histories.^13/

Additional analysis of alternative data offers further encouragement. An analysis of TransUnion credit files by the Political and Economic Research Council found that the inclusion of alternative data—in this case, energy utility and telecommunications payment data—in credit decisions decreased the risk to lenders while increasing access for borrowers. The use of alternative data made it easier for lenders to extend credit, with minorities and low-income individuals benefiting more, in terms of an increase in acceptance rates, than the other borrower subgroups analyzed.^14/

Objections and costs

Early results are positive, but barriers and objections to widespread use of alternative data exist. One of the most formidable barriers is getting reliable data on a large scale. While data furnishers stand to benefit from increased reporting, information sharing presents regulatory and economic hurdles.

In the case of energy utilities and telecommunications service providers, shown in at least one study to be the most promising alternative data source,^15/ there is evidence that when customers know their payment history is being reported, they are more likely to make payments on time. However, the costs of implementing a data reporting system are high. Many data furnishers would face making extensive efforts to consolidate their reporting systems into a central system. These costs may be greater than the perceived benefits. Proponents of the use of alternative data and scoring will have to do more to make a strong business case to potential data furnishers.

Utility providers also face potentially daunting regulatory issues. Several states have laws that prohibit regulated utility companies from sharing customer data. In these states, no customer data from telephone, electric, gas, or water companies can be shared with CRAs. In some other states, it is uncertain whether utility companies are permitted to report data. Advocates for alternative data are urging policymakers to resolve these issues by clarifying whether utility companies can lawfully share consumer data. In many cases, this will require legislative action.^16/

Alternative data reporting also faces objections from consumer advocates. While many advocates recognize and welcome the possibility that alternative scoring would offer more consumers an opportunity to participate in the mainstream credit market, many are concerned that alternative data would be used more often by high-cost lenders to target their services to vulnerable consumers.

Consumer advocates also warn against moving too quickly. There is concern that alternative scores will begin affecting credit access and prices before there are sufficient data to reliably establish the predictive value of scores. A final set of concerns focuses on data sharing. Advocates are concerned that sharing alternative data may violate privacy rights and put consumers at increased risk of identity theft.^17/

The time is now

Proponents of alternative scoring would argue that helping 35 million to 70 million credit-underserved Americans enter the mainstream credit market is worth the expense, effort, and perceived risk. They would also argue that the time for creating a workable alternative is now, because the situation of the credit-underserved may worsen in the future. If the trends toward automation and efficiency that have shaped today's credit market continue, lenders and other businesses will become more dependent on credit scoring and automated underwriting. Consequently, the credit-underserved population will fall further behind in its ability to access affordable credit products and build assets.

If credit markets tighten in response to the subprime mortgage crisis, the challenges for credit-underserved consumers may compound. Assuming that lending standards would rise in a tightened market, lenders would be even less likely to extend credit to no-file or thin-file consumers, no matter how creditworthy those consumers might actually be.

It is difficult to predict exactly how current economic conditions will influence the mainstream credit market, but one trend is likely to continue: Consumers who demonstrate solid credit histories will continue to be well positioned to access credit at a low cost. As the work of proponents and researchers suggests, collecting and scoring meaningful alternative data would position many credit-underserved individuals to reap that same benefit.

For more information The following resources from the Federal Reserve System provide further discussion of credit reporting and scoring. Robert B. Avery, Paul S. Calem, and Glenn B. Canner, "An Overview of Consumer Data and Credit Reporting," Federal Reserve Bulletin, February 2003. Report to the Congress on Credit Scoring and Its Effects on the Availability and Affordability of Credit, Board of Governors of the Federal Reserve System, August 2007. Robert M. Hunt, A Century of Consumer Credit Reporting in America, Federal Reserve Bank of Philadelphia, Working Paper No. 05-13, June 2005.

Understanding FICOs: What's the score? A credit score is a numerical representation of information in an individual consumer's credit report. Credit scores are point-in-time,"snapshot" calculations made when a lender requests a credit report from a Credit Reporting Agency (CRA). Credit scores are fluid; they change over time as the elements in a given credit report change. There are many different credit scores used in the financial services industry. Scores may vary from lender to lender and loan type to loan type (e.g., mortgage loan to auto loan) based on the scoring system used and which CRA's report is the basis for the calculation. Ultimately, the lender decides which score to use. As our main article notes, the most widely used and well known credit score in the United States is the FICO (Fair Isaac Corporation) score. The FICO scoring system bases its prediction of a consumer's future behavior on a comparison between the credit history of the consumer in question and historical profiles of consumers with similar credit histories. For example, a borrower with two 30-days-late payments will be scored against a similar population of borrowers. That borrower will then be graded according to the risk-determining variables used by the scoring system, resulting in a ranking of the borrower within the group of similar borrowers. The FICO score considers five areas of a consumer's credit profile and assigns a relative weight to each. See the chart above for details. FICO scores range from 300 to 850, with a median score of 723.* That means about half of the scores awarded are above this level and half are below it. The higher the FICO score, the more likely a person is to be approved for loans and receive favorable interest rates. For complete information on the FICO score, visit www.myfico.com *Median FICO Score in the U.S. as of March 3, 2008, as reported on www.myfico.com.

1/ Janice Horan, FICO Scores and the Credit Underserved Market, prepared for the Brookings Institution Roundtable on Using Alternative Data Sources in Credit Scoring, December 2005.

2/ Give Credit Where Credit Is Due: Increasing Access to Affordable Mainstream Credit Using Alternative Data, Political and Economic Research Council (PERC) and The Brookings Institution Urban Markets Initiative, 2006.

3/ Ibid.

4/ From the Fair Isaac Corporation's FIC 50 project. See http://image.fiids.com/fi/images/FI50/50anniversaryD.html.

5/ Malgorzata Wozniacka and Snigdha Sen, "Credit Scores—What You Should Know About Your Own," prepared for Secret History of the Credit Card, a project of FRONTLINE and The New York Times, November 2004. See www.pbs.org/wgbh/pages/frontline/shows/credit.

6/ Marvin M. Smith, Ph.D., Recent Developments in Credit Scoring: A Summary, Federal Reserve Bank of Philadelphia, October 2006.

7/ Wozniacka and Sen.

8/ Matt Fellowes, Credit Scores, Reports, and Getting Ahead in America, The Brookings Institution, May 2006.

9/ Giving Underserved Consumers Better Access to the Credit System: The Promise of Non-Traditional Data, Information Policy Institute (IPI), July 2005.

10/ The Fair Credit Reporting Act: Access, Efficiency, and Opportunity, IPI, June 2003.

11/ Horan.

12/ Katy Jacob and Rachel Schneider, Market Interest in Alternative Data Sources and Credit Scoring, Center for Financial Services Innovation (CFSI), December 2006.

13/ Rachel Schneider and Arjan Schütte, "The Predictive Value of Alternative Credit Scores," CFSI, November 2007.

14/ PERC and Brookings, Give Credit Where Credit is Due.

15/ IPI, Giving Underserved Consumers Better Access to the Credit System.

16/ Ibid.

17/ Anna Afshar, "Use of Alternative Credit Data Offers Promise, Raises Issues," New England Community Developments, Federal Reserve Bank of Boston, Issue 1, Third Quarter 2005.